We studied the suitability of collagen-based semisolid medium for assay of endogenous erythroid colony formation performed in myeloproliferative disorders. Bone marrow (BM) mononuclear cells (MNC) from 103 patients suspected of having polycythemia vera (PV, 76 patients) or essential thrombocythemia (ET, 27 patients) were grown in collagen-based, serum-free, cytokine-free semisolid medium. Colony analysis at day 8 or 10 showed that this collagen assay is specific, as endogenous growth of erythroid colonies was never observed in cultures of 16 healthy donors and 6 chronic myelogenous leukemia (CML) patients. Endogenous erythroid colony formation was observed in 53.3% of patients suspected of PV, with only 15.4% of positive cultures for patients with 1 minor PV criterion and 72% (p = 0.009) of positive cultures for patients with greater than or equal to 2 minor or 1 major PV criterion. Similarly, endogenous growth of erythroid colonies was found in 44.4% of patients suspected of ET, with 31.6% of positive cultures for patients with 1 ET criterion versus 75% for patients with greater than or equal to 2 ET criteria. In addition, we found that in collagen gels, tests of erythropoietin (EPO) hypersensitivity in the presence of 0.03. or 0.05 U/ml of EPO and tests of endogenous colony-forming units-megakaryocyte (CFU-MK) formation cannot be: used to detect PV or ET, as these tests were positive for, respectively, 21.4% and 50% of healthy donors and 83% and 50% of CML patients. A retrospective analysis suggests that collagen assays are more sensitive than methylcellulose assays to assess endogenous growth of erythroid colonies. In summary, serum-free collagen-based colony assays are simple and reliable assays of endogenous growth of erythroid colonies in myeloproliferative diseases. They also appear to be more sensitive than methylcellulose-based assays.
